.png){kind=icon}
.png){kind=icon}
Return
Feasibility Study of a Sensorized Prosthetic Foot for Gait Monitoring and Prosthesis Control
.png){kind=icon}
PRE
AI
A
I
F
F
E
N
E
S
DOI：10.1109/JSEN.2026.3691495.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
The assessment of gait parameters in daily life offers valuable insights for tailoring rehabilitation protocols and personalizing lower limb assistive devices. Integrating sensory systems into prostheses may enable adaptive control and continuous gait monitoring. Traditional assessment methods such as motion capture, force platforms (FPs), and wearable inertial measurement units (IMUs) are effective but often lack seamless integration with prostheses, limiting their use in everyday environments. This article presents a novel plantar sensory system designed for integration with lower limb prostheses. The system employs optoelectronic plantar pressure sensors embedded between the prosthetic foot and cosmetic cover at load-bearing regions of the gait cycle. Endurance testing confirmed its robustness over prolonged use. Ten healthy subjects participated in a study evaluating the system’s ability to estimate vertical ground reaction force (vGRF) during walking. A Gaussian process regressor (GPR) was trained on sensory signals with force plate data as ground truth. Using leave-one-subject-out (LOSO) cross-validation, the model showed good generalizability (R<inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">${}^{{2}} = 0.94~\pm ~0.04$ </tex-math></inline-formula>), high correlation with FP data [Pearson r = 0.94 (0.05)], and comparable peak vGRF values. Model stability was assessed by repeating the training <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$30\times $ </tex-math></inline-formula>, yielding a coefficient of variation of the RMSE equal to 7.7%. The trained model was also applied to unseen data from two subjects with transtibial amputation, producing physiologically consistent vGRF profiles. The system also promptly detected gait events, identifying heel strike (HS) and toe-off with median delays of 0.01 and 0.002 s, respectively. These results support the potential of the system for gait parameter estimation in prosthetic applications.
Keywords:
Gait monitoring
optoelectronic sensors
prosthetics
wearable sensors
Journal
IF:
4.5
Papers:
2.1W
Citations:
7.3W
Organization
Cited Papers